中国地震科学实验场人工智能实时地震监测分析系统的应用与展望

doi:10.12196/j.issn.1000-3274.2021.03.001

摘要/Abstract

摘要： “智能地动”(EarthX)系统是目前国际上唯一实时运行的人工智能地震监测系统。 2018年12月起, “智能地动”系统在中国地震科学实验场试运行, 实时处理川滇地区123个地震台站数据。该系统利用人工智能技术自动检测、拾取震相进而定位地震, 在第一个台站接收到P波后数秒内快速产出地震的震源位置和震级, 自发震时刻开始计算, 平均定位仅用时28.9 s。此外, 系统还使用全波形拟合的方法, 在1~3分钟左右快速产出震源机制解和矩震级, 而无须任何人工干预。 EarthX系统在2020年1月1日至2020年8月4日期间, 共记录地震897次, 产出M3.0以上地震震源机制解81次。基本实现了3级以上地震的自动定位和震源机制解的产出, 机制解的平均产出时间为震后103.8 s, 弥补了当前地震监测台网不能产出地震震源机制解的空白。通过EarthX系统的推广应用, 可逐步取代传统的地震监测发布手段, 将台网人员从繁重的地震数据处理工作中解放出来。

关键词: 人工智能, 实时地震监测, 机器学习, 地震搜索引擎, 震源机制

Abstract: The “Artificial Intelligence Earthquake Monitoring” (EarthX) system is currently the only real-time artificial intelligence seismic monitoring system in the world. Since December 2018, the EarthX system has been trial-operated at the China Seismic Experimental Site to process earthquakes recorded by 123 seismic stations in Sichuan and Yunnan in real time. The focal location and magnitude of the earthquakes are quickly shown within a few seconds after the first P waveform of the stations are received, and the EarthX system only takes about 28.9 seconds to locate an earthquake. The focal mechanism solution and moment magnitude of the earthquakes are quickly produced in about 1 to 3 minutes without any manual intervention. During the period from January 1, 2020 to August 4, 2020, the EarthX system recorded a total of 897 earthquakes and produced 81 focal mechanism solutions for earthquakes above M3.0. The automatic locating of earthquakes of magnitude 3 and above and the automatic output of focal mechanism solutions are basically realized in the system and the focal mechanism solutions are produced in average 103.8 seconds after the earthquake occurrence. The EarthX system solves the problem of producing earthquake focal mechanism solutions accurately in a short time. Through the popularization and application of the EarthX system, it can gradually replace the traditional seismic monitoring system, and liberate the seismic analysts from heavy seismic data processing work.

Key words: Artificial intelligence, Real-time seismic monitoring, Machine learning, Earthquake search engine, Focal mechanism

中图分类号:

P315.7

周连庆, 赵翠萍, 张捷, 车时. 中国地震科学实验场人工智能实时地震监测分析系统的应用与展望[J]. 地震, 2021, 41(3): 1-21.

ZHOU Lian-qing, ZHAO Cui-ping, ZHANG Jie, CHE Shi. Application and Prospect of Artificial Intelligence Real-time Seismic Monitoring and Analysis System at the China Seismic Experimental Site[J]. EARTHQUAKE, 2021, 41(3): 1-21.

参考文献

[1] Kong Q, Trugman D T, Ross Z E, et al. Machine learning in seismology: Turning data into insights[J]. Seismological Research Letters, 2019, 90(1): 3-14.
[2] Zhu W, Beroza G C. PhaseNet: A deep-neural-network-based seismic arrival time picking method[J]. Geophysical Journal International, 2018, 216(1): 261-273.
[3] Ross Z E, Meier M-A, Hauksson E. P-Wave Arrival picking and first-motion polarity determination with deep learning[J]. Journal of Geophysical Research: Solid Earth, 2018, 123(6): 5120-5129.
[4] Perol T, Gharbi M, Denolle M. Convolutional neural network for earthquake detection and location[J]. Science Advances, 2018, 4(2): e1700578.
[5] Yoon C E, O’Reilly O, Bergen K J, et al. Earthquake detection through computationally efficient similarity search[J]. Science Advances, 2015, 1(11): e1501057.
[6] 于子叶, 储日升, 盛敏汉. 深度神经网络拾取地震P和S波到时[J]. 地球物理学报, 2018, 61(12): 4873-4886.
YU Zi-ye, CHU Ri-sheng, SHENG Min-han. Pick onset time of P and S phase by deep neural network[J]. Chinese Journal of Geophysics, 2018, 61(12): 4873-4886 (in Chinese).
[7] Wang J, Xiao Z W, Liu C, et al. Deep learning for picking seismic arrival times[J]. Journal of Geophysical Research: Solid Earth, 2019, 124(7): 6612-6624.
[8] Zhu L J, Peng Z G, McClellan J, et al. Deep learning for seismic phase detection and picking in the aftershock zone of 2008 M7.9 Wenchuan earthquake[J]. Physics of the Earth and Planetary Interiors, 2019, 293: 106261.
[9] 蒋一然, 宁杰远. 基于支持向量机的地震体波震相自动识别及到时自动拾取[J]. 地球物理学报, 2019, 62(1): 361-373.
JIANG Yi-ran, NING Jie-yuan. Automatic detection of seismic body-wave phases and determination of their arrival times based on support vector machine[J]. Chinese Journal of Geophysics, 2019, 62(1): 361-373 (in Chinese).
[10] 赵明, 陈石, 房立华, 等. 基于U形卷积神经网络的震相识别与到时拾取方法研究[J]. 地球物理学报, 2019, 62(8): 3034-3042.
ZHAO Ming, CHEN Shi, FANG Li-hua. Earthquake phase arrival auto picking based on U-shaped convolutional neural network[J]. Chinese Journal of Geophysics, 2019, 62(8): 3034-3042 (in Chinese).
[11] Zhang X, Zhang J, Yuan C C, et al. Locating induced earthquakes with a network of seismic stations in Oklahoma via a deep learning method[J]. Scientific Reports, 2020, 10(1): 1941.
[12] Tang L L, Zhang M, Wen L X. Support vector machine classification of seismic events in the Tianshan Orogenic Belt[J]. Journal of Geophysical Research: Solid Earth, 2020, 125(1): e2019JB018132.
[13] 赵明, 陈石, Dave Yuen. 基于深度学习卷积神经网络的地震波形自动分类与识别[J]. 地球物理学报, 2019, 62(1): 374-382.
ZHAO Ming, CHEN Shi, Dave Yuen. Waveform classification and seismic recognition by convolution neural network[J]. Chinese Journal of Geophysics, 2019, 62(1): 374-382 (in Chinese).
[14] Hu J, Qiu H R, Zhang H J, et al. Using deep learning to derive shear-wave velocity models from surface-wave dispersion data[J]. Seismological Research Letters, 2020, 91(3): 1738-1751.
[15] Bianco M J, Gerstoft P, Olsen K B, et al. High-resolution seismic tomography of Long Beach, CA using machine learning[J]. Scientific Reports, 2019, 9(1): 14987.
[16] Zhang J, Zhang H J, Chen E H, et al. Real-time earthquake monitoring using a search engine method[J]. Nature Communications, 2014, 5(1): 5664.
[17] Li Z F, Meier M-A, Hauksson E, et al. Machine learning seismic wave discrimination: Application to earthquake early warning[J]. Geophysical Research Letters, 2018, 45(10): 4773-4779.
[18] Kong Q, Allen R M, Schreier L, et al. MyShake: A smartphone seismic network for earthquake early warning and beyond[J]. Science Advances, 2016, 2(2): e1501055.
[19] Kong Q, Allen R M, Schreier L. MyShake: Initial observations from a global smartphone seismic network[J]. Geophysical Research Letters, 2016, 43(18): 9588-9594.
[20] Khoshnevis N, Taborda R. Prioritizing ground-motion validation metrics using semisupervised and supervised learning[J]. Bulletin of the Seismological Society of America, 2018, 108(4): 2248-2264.
[21] Trugman D T, Shearer P M. Strong correlation between stress drop and peak ground acceleration for recent M 1-4 earthquakes in the San Francisco Bay Area[J]. Bulletin of the Seismological Society of America, 2018, 108(2): 929-945.
[22] Zhu H, Sun Y, Zhao W, et al. Rapid learning of earthquake felt area and intensity distribution with real-time search engine queries[J]. Scientific Reports, 2020, 10(1): 5437.
[23] DeVries P M R, Viégas F, Wattenberg M, et al. Deep learning of aftershock patterns following large earthquakes[J]. Nature, 2018, 560(7720): 632-634.
[24] Rouet-Leduc B, Hulbert C, Lubbers N, et al. Machine learning predicts laboratory earthquakes[J]. Geophysical Research Letters, 2017, 44(18): 9276-9282.
[25] Peng Z, Zhao P. Migration of early aftershocks following the 2004 Parkfield earthquake[J]. Nature Geoscience, 2009, 2(12): 877-881.
[26] Zhang M, Wen L X. An effective method for small event detection: match and locate (M&L)[J]. Geophysical Journal International, 2015, 200(3): 1523-1537.
[27] Beaucé E, Frank W B, Romanenko A. Fast matched filter (FMF): An efficient seismic matched-filter search for both CPU and GPU architectures[J]. Seismological Research Letters, 2018, 89(1): 165-172.
[28] Ross Z E, Trugman D T, Hauksson E, et al. Searching for hidden earthquakes in Southern California[J]. Science, 2019, 364(6442): 767-771.
[29] Liu M, Zhang M, Zhu W Q, et al. Rapid characterization of the July 2019 Ridgecrest, California, earthquake sequence from raw seismic data using machine-learning phase picker[J]. Geophysical Research Letters, 2020, 47(4): e2019GL086189.
[30] Pesicek J D, Thurber C H, Zhang H, et al. Teleseismic double-difference relocation of earthquakes along the Sumatra-Andaman subduction zone using a 3-D model[J]. Journal of Geophysical Research: Solid Earth, 2010, 115(B10): B10303.
[31] Waldhauser F. Near-Real-Time double-difference event location using Long-Term seismic archives, with application to northern California[J]. Bulletin of the Seismological Society of America, 2009, 99(5): 2736-2748.
[32] Waldhauser F, Ellsworth W L. Fault structure and mechanics of the Hayward Fault, California, from double-difference earthquake locations[J]. Journal of Geophysical Research: Solid Earth, 2002, 107(B3): 2054.
[33] Allen R. Automatic phase pickers-their present use and future-prospects[J]. Bulletin of the Seismological Society of America, 1982, 72(6): S225-S242.
[34] Allen R. Automatic earthquake recognition and timing from single traces[J]. Bulletin of the Seismological Society of America, 1978, 68(5): 1521-1532.
[35] Leonard M, Kennett B L N. Multi-component autoregressive techniques for the analysis of seismograms[J]. Physics of the Earth and Planetary Interiors, 1999, 113(1-4): 247-263.
[36] Sleeman R, van Eck T. Robust automatic P-phase picking: an on-line implementation in the analysis of broadband seismogram recordings[J]. Physics of the Earth and Planetary Interiors, 1999, 113(1-4): 265-275.
[37] 罗钧, 赵翠萍, 周连庆. 川滇块体及周边区域现今震源机制和应力场特征[J]. 地震地质, 2014, 36(2): 405-421.
LUO Jun, ZHAO Cui-ping, ZHOU LIAN-qing. Characteristics of focal mechanisms and stress field of the Chuan-Dian rhombic block and its adjacent regions[J]. Seismology and Geology, 2014, 36(2): 405-421 (in Chinese).
[38] Lei X L, Su J R, Wang Z W. Growing seismicity in the Sichuan Basin and its association with industrial activities[J]. Science China Earth Sciences, 2020, 63(11): 1633-1660.
[39] Richter C F. An instrumental earthquake magnitude scale[J]. Bulletin of the Seismological Society of America, 1935, 25(1): 1-32.
[40] Gutenberg B. Amplitudes of surface waves and magnitudes of shallow earthquakes[J]. Bulletin of the Seismological Society of America, 1945, 35(1): 3-12.
[41] Gutenberg B. Magnitude determination for deep-focus earthquakes[J]. Bulletin of the Seismological Society of America, 1945, 35(3): 117-130.
[42] Gutenberg B. Amplitudes of P, PP, and S and magnitude of shallow earthquakes[J]. Bulletin of the Seismological Society of America, 1945, 35(2): 57-69.
[43] Gutenberg B, Richter C F. Magnitude and energy of earthquakes[J]. Annali di Geofisica, American Association for the Advancement of Science, 1956, 9(1): 6-12.
[44] Scordilis E M. Empirical global relations converting M_S and m_b to moment magnitude[J]. Journal of Seismology, 2006, 10(2): 225-236.
[45] Kanamori H. The energy release in great earthquakes[J]. Journal of Geophysical Research, 1977, 82(20): 2981-2987.
[46] Hanks T C, Kanamori H. A moment magnitude scale[J]. Journal of Geophysical Research: Solid Earth, 1979, 84(B5): 2348-2350.
[47] 陈章立, 陈翰林, 赵翠萍, 等. 地震大小的度量[J]. 地震, 2014, 34(1): 1-12.
CHEN Zhang-li, CHEN Han-lin, ZHAO Cui-ping, et al. Measuring the earthquake size[J]. Earthquake, 2014, 34(1): 1-12 (in Chinese).
[48] Gupta H K. The present status of reservoir induced seismicity investigations with special emphasis on Koyna earthquakes[J]. Tectonophysics, 1985, 118(3-4): 257-279.
[49] Gupta H K. A review of recent studies of triggered earthquakes by artificial water reservoirs with special emphasis on earthquakes in Koyna, India[J]. Earth-Science Reviews, 2002, 58(3-4): 279-310.
[50] Walsh F R, Zoback M D. Oklahoma’s recent earthquakes and saltwater disposal[J]. Science Advances, 2015, 1(5): e1500195.
[51] Foulger G R, Wilson M P, Gluyas J G, et al. Global review of human-induced earthquakes[J]. Earth-Science Reviews, 2018, 178: 438-514.
[52] Grigoli F, Cesca S, Priolo E, et al. Current challenges in monitoring, discrimination, and management of induced seismicity related to underground industrial activities: A European perspective: Challenges in induced seismicity[J]. Reviews of Geophysics, 2017, 55(2): 310-340.
[53] 庄建仓, 马丽. 主震和余震从大森公式到ETAS模型[J]. 国际地震动态, 2000(5): 12-18.
ZUANG Jian-cang, MA Li. Main shock and after shock: From the Omori formula to the ETAS model[J]. Recent Development in World Seismology, 2000(5): 12-18 (in Chinese).
[54] 宋金, 蒋海昆. 序列衰减与余震激发研究进展及应用成果[J]. 地震地质, 2009, 31(3): 559-571.
SONG Jin, JIANG Hai-kun. A Review on Decay and Generation of Aftershock Activity[J]. Seismology and Geology, 2009, 31(3): 559-571 (in Chinese).
[55] 崔庆谷, 林国良, 李倩. 利用近震源区强震记录识别鲁甸6.5级地震后叠加的余震事件[J]. 大地测量与地球动力学, 2015, 35(6): 1069-1073.
CUI Qing-gu, LIN Guo-liang, LI Qian. Recognizing aftershocks merged in main shock’s coda wave of Ludian M6.5 by strong acceleration recording near source region[J]. Journal of Geodesy and Geodynamics, 2015, 35(6): 1069-1073 (in Chinese).
[56] 黄媛, 吴建平, 张天中, 等. 汶川8.0级大地震及其余震序列重定位研究[J]. 中国科学(D辑), 2008, 38(10): 1242-1249.
HUANG Yuan, WU Jian-ping, ZHANG Tian-zhong. Study on the relocation of Wenchuan M8.0 earthquake and its aftershock sequence[J]. Science in China (Series D), 2008, 38(10): 1242-1249 (in Chinese).
[57] 徐锡伟, 闻学泽, 叶建青, 等. 汶川M_S8.0地震地表破裂带及其发震构造[J]. 地震地质, 2008, 30(3): 597-629.
XU Xi-wei, WEN Xue-ze, YE Jian-qing, et al. The M_S8.0 Wenchuan earthquake surface ruptures and its seismogenic structure[J]. Seismology and Geology, 2008, 30(3): 597-629 (in Chinese).
[58] Allen R M, Gasparini P, Kamigaichi O, et al. The status of earthquake early warning around the world: An introductory overview[J]. Seismological Research Letters, 2009, 80(5): 682-693.